A Prediction Method of Airport Noise Based on Hybrid Ensemble Learning

• Main Authors: Tao XU, Qichuan YANG, Zonglei LV
• Format: Article
• Language: English
• Published: IFSA Publishing, S.L. 2014-05-01
• Series: Sensors & Transducers
• Subjects: Airport noise; Hybrid ensemble; Artificial neural network; Nearest neighbor; Nonlinear regression.
• Online Access: http://www.sensorsportal.com/HTML/DIGEST/may_2014/Vol_171/P_RP_0127.pdf

Using monitoring history data to build and to train a prediction model for airport noise is a normal method in recent years. However, the single model built in different ways has various performances in the storage, efficiency and accuracy. In order to predict the noise accurately in some complex environment around airport, this paper presents a prediction method based on hybrid ensemble learning. The proposed method ensembles three algorithms: artificial neural network as an active learner, nearest neighbor as a passive leaner and nonlinear regression as a synthesized learner. The experimental results show that the three learners can meet forecast demands respectively in on- line, near-line and off-line. And the accuracy of prediction is improved by integrating these three learners’ results.